Help
Comprehensive guide to using DRIFT for drug target identification through compound similarity.
What is DRIFT
Only a small portion of small molecules bind to the expected targets, while side effects of marketed drugs are frequently caused by off-target binding. Identifying the protein targets to which a compound may bind is crucial for early stage of drug discovery, such as drug toxicity evaluation and drug repurposing. A widely accepted assumption in drug discovery is that chemical compounds with similar structures tend to have similar biological activity, particularly protein binding profiles. Thus unknown small molecule targets can be predicted by exploring structurally similar bait compounds with annotated targets.
DRIFT (Drug Repositioning and Identification via Fingerprint-based Target inference) is a computational tool for predicting protein targets for an unknown chemical compound by comparing it against curated libraries of bioactive molecules with known target associations. The web server provides fast query entry, support for name, SMILES, or file-based compound input, flexible similarity search configurations, and ranked target evidence drawn from databases such as ChEMBL, ZINC, HMDB, and BindingDB.
Submit a query compound
Search the compound
You can search for a compound by name (e.g. aspirin), molecular formula (e.g. C9H8O4), SMILES string (e.g. C1=CC=C(C=C1)C=O), or database identifiers such as CHEMBL ID, HMDB ID, or ZINC ID. Enter your query into the search box and press Enter or click the Search button. DRIFT will resolve the compound identity from its internal databases and proceed with the similarity search.
Upload the compound file
You can upload a structure file in .sdf or .txt format. Click on “Upload Structure” on the homepage or select the File tab on the submit page, then choose a file from your computer. DRIFT will parse the compound structure from the file and use it for the similarity search.
Draw the compound structure
You can draw a compound structure directly using the built-in molecular editor. Click on “Draw Structure” on the homepage or select the Draw tab on the submit page. The structure editor allows you to sketch a molecule, which will be converted to a SMILES string and submitted for target prediction.
Parameters to set
Database
DRIFT supports searching against multiple curated databases: CHEMBL, HMDB, ZINC, and BindingDB. CHEMBL is the default and most comprehensive option, containing a large collection of bioactive compounds with known target annotations. You can select a different database depending on your research needs.
Similarity Cutoff
The similarity cutoff determines how structurally similar a bait compound must be to the query compound to be included in the target inference. The default value is 0.1, using the FP2 fingerprint (a path-based fingerprint that encodes linear and branched atom connectivity). A lower cutoff (closer to 0) includes more diverse results, while a higher cutoff restricts results to compounds that are more structurally similar to the query.
Number of compounds
This parameter controls the maximum number of structurally similar bait compounds to retrieve and use for target aggregation. The default value is 100. Increasing this number may capture more potential targets at the cost of longer processing time.
Submit multiple compounds in batch mode
DRIFT supports batch submission of multiple compounds for efficient high-throughput screening. In batch mode, you can upload a file containing multiple compounds (one per line as SMILES, or in SDF format) or paste multiple identifiers. Each compound is processed independently, and results are aggregated into a single queue entry that you can inspect after processing completes.
Contact
If you encounter any problems or have questions about DRIFT, please contact us at [email protected] (CC: [email protected]).